An ensemble learning-based model for effective chronic kidney disease prediction

Due to rising chronic kidney disease (CKD) cases across the globe, it is required to be detected and diagnosed effectively. Machine learning-based models can be an effective tool for early and effective predictions. In this work, we apply machine, deep, and ensemble learning-based models to predict the cases of CKD and non-CKD. We have also applied the stacking and voting ensemble of SVM, RF, Adaboost, LDA, and MLP models to effectively and quickly predict the CKD and non-CKD cases. Accuracy, Precision, Recall, F1-Score, and ROC are used to evaluate the models’ performance. Further, the Freidman ranking is also calculated to measure the effectiveness of the models. A comparative study concludes that ensemble learning-based models are the best-performing models.